Normal human c-fgr cDNA clones were constructed by using normal peripheral blood mononuclear cell mRNA as a template. Nucleotide sequence analysis of two such clones revealed a 1,587-base-pair- long open reading frame which predicted the primary amino acid sequence of the c-fgr translational product. Homology of this protein with the v-fgr translational product stretched from codons 128 to 516, where 32 differences among 388 codons were observed. Sequence similarity with human c-src, c-yes and fyn translational products began at amino acid position 76 of the predicted c-fgr protein and extended nearly to its C-terminus. In contrast, the stretch of 75 amino acids at the N-terminus demonstrated a greatly reduced degree of relatedness to these same proteins. To verify the deduced amino acid sequence, antibodies were prepared against peptides representing amino- and carboxy-terminal regions of the predicted c-fgr translational product. Both antibodies specifically recognized a 55-kilodalton protein expressed in COS- 1 cells transfected with a c-fgr cDNA expression plasmid. Moreover, the same protein was immuno-precipitated from an Epstein- Barr virus-infected Burkitt's lymphoma cell line which expressed c-fgr mRNA but not in its uninfected fgr mRNA-negative counterpart. These findings identified the 55-kilodalton protein as the product of the human fgr proto-oncogene.